An application management strategy can help IT maintain effective performance and service levels and prevent costly failures. Here are some key features and capabilities to look for in an application management solution.
Today's IT organizations are challenged to reduce operational expenses and overhead, while still improving service levels and maintaining quality. When business-critical applications go down, the failure affects IT on all of these fronts and can have a significant impact on the company's bottom line.
The shift from basic infrastructure management to service management — where application issues are aligned with, and prioritized by, their relevance to the key business processes — has greatly increased the importance of combining visibility into application transactions, end-user experiences, and delivered service levels.
As new technologies and business demands evolve, IT must have a true application management strategy for application performance and service levels. Without it, the IT staff will continue to face the following key challenges with their performance management tools:
- How to align IT with business objectives
- How to deal and cope with dynamic changes in the data center
- How to manage complex object-oriented and service-oriented architecture (SOA) applications
- How to increase communication in the IT organization across team structures
- How to constrain the cost of monitoring and management
Here are 10 requirements for application management tools that will address these visibility challenges and help avoid expensive application downtime.
Note: This article is also available as a PDF download.
1: Monitor more than just infrastructure components
The only reason technical issues matter is that they affect business services. As IT begins to move closer toward alignment with business objectives, it's important to monitor your environment from the point of view of how infrastructure problems affect your applications, services, the end-user experience, and, therefore, your business objectives.
2: Adapt and change as your service delivery maturity level continues to improve
Role-based dashboards, for example, should automatically adapt to infrastructure changes, and a flexible template-based architecture should allow you to deploy your application management solution in incremental steps to reduce time to value.
3: Manage the dynamic change that comes with a virtualized environment
Your application management solution of choice must be able to monitor and follow your applications as they dynamically move from one virtual machine to another. It's the applications that give meaning to, and determine the importance of, monitoring data. In other words, you need to be able to understand not just the virtual infrastructure, but also the relationships and interactions between all its components. This helps administrators detect, diagnose, and resolve incidents and problems, ensure performance, and perform capacity planning — all of which become more complex in virtual environments.
4: Correlate change to the root cause of incidents across the stack in a virtual environment
Your toolset should help you understand the way change affects performance across the stacks. You must be able to correlate change with performance — regardless of all the different components lurking around in your virtual environment — because once you have virtualized, the traditional association between a technical component and an application or business service no longer exists.
5: Treat a single component in multiple ways
As applications and services share components, your application management toolset must be able to prioritize and manage the availability and performance of individual components based on the business priority of all the applications and services using them.
6: Administer and adapt to tooling changes quickly
Your application management toolset must be able to dynamically discover changes to your underlying architecture and automatically update key metrics, scripts, and dashboards as new services and composite applications are brought online. Extensive administration every time something changes is not cost effective and reduces IT nimbleness.
7: Allow teams to use their existing tools as needed
Although a completely homogenous monitoring environment has its attractions, it is not always easy to square that with ensuring that each team can do its job as effectively as possible. Different internal and external teams often have their own monitoring tools and, in some cases, may need to continue using them. Therefore, it is important that your overall application management solution be able to integrate with other toolsets as appropriate.
8: Help specialist teams collaborate with others
Most organizations have domain-specific IT teams based on individual technologies. Yet business services cut across those technologies. So while performance management involves collecting data, it is also about enabling collaboration between those different teams — each of which may have different perspectives and technical skill sets. Your performance toolset should be a vehicle for cross-team collaboration, providing both high-level overviews and specific and relevant informational views for each team. This minimizes the time spent by domain teams on issues that, while relevant to their own technical world, are not highly business-relevant.
9: Reduce the cost per metric of monitoring
Although greater efficiencies and improved technologies have lowered the cost of monitoring individual metrics over time, the number of metrics that need to be monitored in an average environment has increased exponentially. When data is provided without its associated context, the tendency is to collect it all and try to figure it out later. This is a poor strategy. In fact, the value of monitoring tools is challenged by some customers because of their inability to cope with the vast amount of data they're collecting. Instead, your toolset must be able to help you focus your deepest data collection in the areas that are most business-valuable.
10: Automate the process of managing metrics, including the way end users are using the application
Today's IT organization needs better tools than it did when monitoring involved just CPUs, disks, and other infrastructure components. Instead, we now need a way to understand all the moving parts of our applications — but in the context of the business service that is being delivered. Automation is the key to making performance monitoring both adaptive and cost effective.
Hugh McEvoy oversees product management for Foglight's applications offerings, which include Java, .NET, ERP, and CRM performance management solutions. His role includes market strategy, technical strategy, and setting development priorities. Hugh joined Quest Software in 1999.